The business world has discovered the Internet in a big way. The Internet was a 25-year overnight success. Even though it has been around since the late 1960's, it wasn't until the mid 90s, with the advent of the graphical web browser, that the Internet took off like a rocket. Since then, businesses started scrambling for ways to send database information around the globe from servers to browser clients. Wait a minute! A centralized server sending data to a client connected to it? This sounds like the 60's all over again, right? Centralized main frame computers talking to light clients. The world is migrating to mainframe servers serving client browsers connected to them. Information centralized on mainframe servers. Hmmm. Where have we heard this before?
So now programmers need to not only access relational data sources, but non-relational data as well. As we mentioned, Microsoft's approach is to provide a common method to get at data stored in various formats. They think it makes sense to focus on the access to the data rather than to the physical layout of the database itself. After all, what if we need to get at data in a relational database, a legacy system, an Excel spreadsheet, a web site, some text files, and e-mail? And what if these are stored in various locations? Rather than change the world to conform to a single data structure, we want to change the way we retrieve data stored in various structures. Makes sense.